[
https://issues.apache.org/jira/browse/CASSANDRA-10249?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jim Plush updated CASSANDRA-10249:
----------------------------------
Comment: was deleted
(was: Uploading some testing screenshots I was doing the last couple days when
trying to establish some benchmarks. With compression off I was looking to do
1million writes (RF3) with 50K reads on a 60 node cluster. with the default of
64K buffer size I/O was saturated and read latency was 100+ms. with the buffer
at 4K I/O was quite stable at that rate. This was a straight row key look up
test. e.g. no wide rows. It was reading way too much data for the queries.
Would it be possible to have the buffer size set on a per table setting?
(screenshots attached))
> Make buffered read size configurable
> ------------------------------------
>
> Key: CASSANDRA-10249
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10249
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Albert P Tobey
> Fix For: 2.1.x
>
> Attachments: Screenshot 2015-09-11 09.32.04.png, Screenshot
> 2015-09-11 09.34.10.png, patched-2.1.9-dstat-lvn10.png,
> stock-2.1.9-dstat-lvn10.png, yourkit-screenshot.png
>
>
> On read workloads, Cassandra 2.1 reads drastically more data than it emits
> over the network. This causes problems throughput the system by wasting disk
> IO and causing unnecessary GC.
> I have reproduce the issue on clusters and locally with a single instance.
> The only requirement to reproduce the issue is enough data to blow through
> the page cache. The default schema and data size with cassandra-stress is
> sufficient for exposing the issue.
> With stock 2.1.9 I regularly observed anywhere from 300:1 to 500
> disk:network ratio. That is to say, for 1MB/s of network IO, Cassandra was
> doing 300-500MB/s of disk reads, saturating the drive.
> After applying this patch for standard IO mode
> https://gist.github.com/tobert/10c307cf3709a585a7cf the ratio fell to around
> 100:1 on my local test rig. Latency improved considerably and GC became a lot
> less frequent.
> I tested with 512 byte reads as well, but got the same performance, which
> makes sense since all HDD and SSD made in the last few years have a 4K block
> size (many of them lie and say 512).
> I'm re-running the numbers now and will post them tomorrow.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
